UsePod

Frontier models are temporary monopolies; open-weight models are commodities, and commodities get cheaper when markets are allowed to work. UsePod is a two-sided marketplace that lets supply meet demand directly and drives the price of inference toward the cost of production, while a drop-in API keeps your existing agents and tools running unchanged. Why inference is becoming a commodity →

Who it’s for

Inference users

Change one base URL and reach frontier models plus open-weight alternatives at marketplace prices: frontier quality where you need it, commodity pricing everywhere else. Quickstart →

Hardware providers

Run open models on your own GPUs, set your prices, and earn 80% of every settled inference in USDC. Run a provider →

Key & credit holders

Hold an upstream API key or idle inference allocation? Connect it and resell into the market at a price you set. Resell your keys →

Researchers

Monetize custom models and quantizations the public APIs don’t offer, priced at whatever the market will bear. Run a provider →

What you get

Drop-in API

OpenAI- and Anthropic-compatible. One base-URL change, with no SDK or auth changes. Drop-in API →

Best-price routing

Candidates are capped at the cheapest centralized price, then filtered by your ceiling, provider health, and capacity. Routing & matching →

Always-on fallback

Centralized providers stay available as a tier-zero fallback, so you’re never stranded. Pricing →

Built for agents

Every page is raw Markdown (append .md), with /llms.txt, /llms-full.txt, and /api/pages.json. Docs for agents →

Start here

Quickstart Get a token, fund it, and send your first request in under a minute.

Pay per request with x402 No account: pay each call in USDC or SOL on Solana, verified on-chain.

Run a provider Install the agent on a GPU, or connect an upstream key, and start earning USDC.

Why inference is a commodity The pricing thesis behind the marketplace.

Verification is the product Where this goes: verifiable compute and a futures market.